Watch a Manchester United match and you will see a word across the players’ chests that most fans could not explain. Snapdragon. It is not a bank, an airline or a betting site.

It is a family of processors, sold to phone and laptop makers, that almost nobody buys directly. Yet its owner paid to put it on one of the most photographed shirts in sport.

That shirt captures the strangest contest in technology marketing. Chipmaker branding is the practice of giving a semiconductor company or its processors a name, identity and reputation that buyers recognise, whether the buyer is a phone maker’s procurement team, a data-centre architect or a teenager choosing a gaming laptop.

Two audiences. Usually very different. The engineers who design a chip into a product care about performance per watt, supply security and roadmap credibility. The person who eventually buys the device mostly cares that it works.

Every chipmaker has to decide how much effort to spend on the second group. Their answers could hardly be more different.

Why would a chip company want fans at all?

Because whoever owns the customer’s attention owns part of the price. A component nobody asks for is a commodity, and commodities get squeezed by the brands that sit on top of them. A component buyers request by name gives its maker bargaining power with the device brand.

Intel proved this in 1991 with Intel Inside, the campaign that turned an invisible part into a household name and paid PC makers to display its logo. Every ingredient-branding play since has been a variation on that move. The logic is simple to state and hard to execute: a hidden part becomes worth naming when end buyers cannot judge its quality for themselves and still feel the cost of choosing badly.

The catch is that a consumer brand creates expectations a chip roadmap cannot always meet. Fame cuts both ways. When the product slips, everyone notices.

Intel: the brand that taught us to ask, then lost the thread

For two decades Intel’s naming was a model of usefulness. Core i3, i5, i7 and later i9 gave shoppers a ladder they could climb without understanding clock speeds. Good, better, best. A salesperson in Sim Lim Square could explain it in ten seconds.

Then the ladder got crowded. Generation numbers piled up, suffix letters multiplied, and enthusiasts joked that each new generation was the previous one with a fresh logo.

When Intel did redraw its logo in 2020, the flatter wordmark struck some as more cosmetics counter than engineering company. Others liked it. Nobody was talking about the chips.

Intel’s own film on the 2020 brand transformation, including the flatter wordmark that divided opinion.

In June 2023 Intel tried a reset. It dropped the “i” from Core, stopped advertising generation numbers and introduced Core Ultra for its strongest processors, launched alongside its first client chips with a dedicated AI engine. Reviewers carried on using generation names regardless, because buyers needed them to tell new from old. When your customers rebuild the naming scheme you just discarded, the scheme was doing a job.

The deeper problem was not the labels. Intel lost manufacturing leadership, and no brand could outrun that. In November 2020 Apple began replacing Intel processors in its Macs with its own M1 chip, a public demonstration that a famous customer could live without the famous ingredient.

The rest reads like a medical chart. Intel announced 15,000 job cuts in August 2024. Pat Gelsinger left as chief executive that December, and Lip-Bu Tan took over in March 2025. By August 2025 the United States government held a 9.9% equity stake, and in September Nvidia invested US$5 billion in its old rival.

Interbrand’s 2025 league table recorded the cost in brand terms: Intel fell to 71st, its value down 42%. Thirty years of “Inside” could not protect a brand whose product story had stalled.

AMD and the underdog that stopped apologising

AMD spent much of its early life as the other name on the box. In 1982 it became a licensed second source of Intel’s x86 processors, because IBM wanted a backup supplier for its PC. That origin shaped how buyers saw it for years: the cheaper alternative, the value pick, the brand you defended in forums.

There were moments of glory. The Athlon arrived in 1999 and gave enthusiasts a reason to look twice. In 2006 AMD bought ATI Technologies, and with it the Radeon graphics line. In 2010 it retired the ATI name and put everything under AMD, a classic master-brand consolidation that traded a well-loved gamer badge for one stronger corporate name.

It also made a structural choice that later looked prescient. In 2008 it announced it would spin off its factories into GlobalFoundries, backed by investors from Abu Dhabi, who had bought out AMD’s final stake by 2009. AMD became a designer of chips, not a maker of them, free to use whichever foundry was ahead.

The turnaround came with product, not posters. Lisa Su became chief executive in October 2014. Ryzen, built on the new Zen architecture, launched in March 2017 and borrowed Intel’s own ladder logic with Ryzen 3, 5 and 7. In 2022 AMD overtook Intel by market value, and in 2023 it phased out the Xilinx brand, acquired the year before.

Notice the pattern. Each time AMD bought something, it absorbed the name. One brand, many products.

Snapdragon: when the product brand outgrows the parent

Qualcomm made the opposite choice. It kept its corporate name for the world of patents, licences and modems, and built Snapdragon as the face the public sees.

The line was announced in late 2006, with the first chips released in 2007, and the name was reportedly picked because it sounded fast and fierce. From 2013 Qualcomm sorted the range into 200, 400, 600 and 800 tiers, giving reviewers and shoppers a ladder of their own. In 2015 it pulled its standalone modems under the Snapdragon name too.

From 2017 it pushed into Windows laptops with Microsoft, which is exactly where a recognisable consumer badge starts to pay. Then came the football. Snapdragon replaced TeamViewer as Manchester United’s main shirt sponsor for the 2024–25 season, in a deal since extended to 2029. The name also sits on Snapdragon Stadium in San Diego.

Snapdragon’s own film on its Manchester United partnership, the shirt deal that put a component brand in front of football fans.

Why spend that money on a component? Because Qualcomm is trying to do in laptops what Intel did in the 1990s: make the chip a reason to choose one machine over another. A shirt reaches a global audience that will never read a benchmark. In 2025 Qualcomm entered Interbrand’s global ranking for the first time, at 39th.

The risk is subtle. When the product brand is bigger than the company, a bad product year damages the only name consumers know.

Nvidia: a gamer’s badge that became the AI economy’s logo

Nvidia was founded in April 1993 by Jensen Huang, Chris Malachowsky and Curtis Priem. The name draws on invidia, Latin for envy, which is a bold thing to call a start-up with about US$40,000 in the bank.

Its smartest branding move came in 1999, when the GeForce 256 became the first product it explicitly marketed as a GPU. Popularising the category word is one of the most powerful things a technology brand can do. If customers describe their need in your vocabulary, you have half won the sale.

For years GeForce belonged to gamers. Then researchers found that graphics processors were superb at training neural networks. Google Brain was using Nvidia hardware in 2009, and in 2016 Nvidia handed the first DGX-1 system to a young lab called OpenAI. When generative AI exploded, the company that owned the hardware, the software tools and the vocabulary was standing in exactly the right place.

The milestones came fast: a US$1 trillion market value in May 2023, the first close above US$4 trillion in July 2025, and the first company to reach US$5 trillion that October. Interbrand’s 2025 ranking moved Nvidia from 36th to 15th, its brand value up 116%, the largest jump in that table’s history.

Yet scale and ecosystem dominance are not, on their own, a brand strategy. Nvidia’s brand today rests on being indispensable. Indispensability is a position, not a personality. Positions can be attacked.

Can a chip company be famous without a consumer brand?

Absolutely. Two of the most important names in the industry barely market themselves to the public, and they are stronger for it.

TSMC was founded in Taiwan in 1987 as the world’s first dedicated semiconductor foundry. It makes chips designed by others, including Apple, Nvidia, AMD, Qualcomm and MediaTek, and holds roughly 70% of the global foundry market. Its logo is a modest lowercase wordmark.

Its real brand promise is to customers: it will deliver, it will keep their designs secret, and it has no chips of its own to sell against them. That promise now travels. Its Kumamoto plant, a joint venture with Sony and Denso, began commercial production in Japan in December 2024, and it is building a cluster of fabs in Arizona.

Arm takes discretion further. It began in 1990 as a joint venture between Acorn, Apple and VLSI Technology, and it does not sell chips at all. It licenses designs, and its architecture sits in the vast majority of the world’s phones. In 2017 it changed its name from ARM to Arm, with a lowercase logo: a small signal of a company growing from acronym to brand.

Arm’s true asset is neutrality. When Nvidia agreed in 2020 to buy Arm from SoftBank for US$40 billion, the deal ran into regulatory and competition objections and collapsed in February 2022. Arm floated on Nasdaq in September 2023 instead. Whatever the legal arguments, the episode showed how much the industry valued a supplier that belongs to no single rival.

For B2B brands, that is the lesson worth writing down. Trust can be the whole brand.

The ghost of Motorola, and the rise of MediaTek

Brands in this industry can vanish faster than their products. Motorola’s 68000 processor, introduced in 1979, powered the original Macintosh 128K, the Amiga, the Atari ST and Sega’s Mega Drive. In 2004 Motorola spun off its semiconductor business as Freescale, and in 2015 Freescale merged into NXP.

Three names in eleven years. Whatever recognition the chips had earned was scattered.

MediaTek, founded in Hsinchu in 1997 as a spin-off from UMC, shows the opposite path. In 2014 it launched an “Everyday Genius” platform aimed at what it called the super-mid market. It then built named ranges, Helio from 2015 and Dimensity for 5G from late 2019, so phone buyers and reviewers had something to compare. By 2020 it was the largest smartphone chipset supplier by volume.

The lesson for any Asian manufacturer is clear enough. Start as the dependable supplier nobody mentions, then add a consumer-facing range name once your customers’ customers begin to care.

What should a business that isn’t a chipmaker learn from this?

Most readers do not design processors. But plenty of Singapore firms sell a component, a service or a platform that disappears inside somebody else’s brand. Picture a precision-engineering supplier in Tuas whose parts sit inside medical devices sold across Europe. Should it brand its parts or stay invisible?

Name your primary audience. If the people who choose you are engineers and buyers, your brand is reliability, documentation and the confidence of your account team; TSMC and Arm are the models. If end users can be persuaded to ask for you, a consumer-facing name may earn its cost, as Snapdragon is betting.

Decide which name carries the weight. AMD concentrated on one master brand. Qualcomm split corporate and product. Both work. Doing a bit of each, with no rule, does not.

Build a naming system people can use. Intel’s i3-to-i9 ladder was valuable because a shop assistant could explain it. Before you change a scheme, check you are not deleting information buyers rely on.

Guard your neutrality if your customers compete with each other. It is easy to lose and almost impossible to buy back.

And no amount of identity work rescues a product that has fallen behind. Intel’s decade is as clear a proof as modern branding offers: reputation follows engineering, not the other way round.

Quick answers from the server room

What is ingredient branding in the chip industry?
Ingredient branding is when a component maker markets its part directly to end buyers so they request it by name. Intel Inside, launched in 1991, is the best-known example, and Qualcomm’s promotion of Snapdragon follows the same logic.

Why doesn’t TSMC advertise to consumers?
TSMC manufactures chips designed by other companies, such as Apple, Nvidia and AMD, so its brand depends on reliability and discretion with those customers rather than public fame.

Did Intel drop the Core i5 and i7 names?
Yes. In June 2023 Intel removed the “i” from its Core branding, introduced Core Ultra for its top chips and stopped promoting generation numbers.

If your brand lives inside somebody else’s product and you are deciding whether to step into the light, talk to Firefish about brand strategy.